Research On Gas Path Fault Diagnosis Technology For Marine Gas Turbine

Marine gas turbine is long-time working in the high-speed, high-temperature,high-pressure and high-stress state conditions, as well as high-humidity and high-saltenvironment, making the possibility of various fault occurs during operation increases, it willbring great losses to operation in the event of failure. Gas turbine fault diagnosis can helppeople of operation and maintenance advance which is abnormal. We should identify thecause of the malfunction quickly and give a reasonable solution, also, make the gas turbinework safely and steadily. In this work, this article researches on technology of gas path faultdiagnosis of triaxial gas turbine for the study of established mathematical modeling,failuremechanism analysis and diagnosis algorithm. The main work is following:(1) Establishment the typical gas path fault mathematical model of gas turbine.According to the relationship of thermal parameters from the gas turbine gas pathcomponents, application of small deviation method makes gas path componentsthermodynamic formula linear method, isolated fault occurs independent performanceparameters change of the whole change, measured the deviation of the parameters and toestablish the fault coefficient matrix the deviation of the performance parameter linked;through experimental data, solving the fault coefficient matrix, and ultimately established thefault between the model matrix in the performance parameters of the measurementparameters of the operating point, the model can be applied to the simulation of the gasturbine failure.(2) Simulated typical gas path failure of gas turbine. Criterion to determine the typicalgas turbine failure; analysed the gas turbine typical failure mechanism, explained the causesof the incident and the fault for the overall performance of the gas turbine; fault model matrixbased on the design point, different gas turbine design operating point of the compressor/fouling of turbine blades, compressor blade tip clearance, a compressor/turbine blades wear,compressor/turbine blades mechanical failure, the hot corrosion of turbine blades andcombustion chamber failure of six typical fault simulation. Made the analysis of thesimulation results combined with gas turbine failure mechanism. The analysis results showthat the fault simulation results obtained are reliable and can be used directly as a faultknowledge for the development of fault diagnosis algorithm.(3) Research on the gas path fault diagnosis technology based on probabilistic neural network. First proposed the idea of a gas turbine fault diagnosis; probabilistic neural networkdiagnostic model followed by the establishment, the application of simulated fault data, thegas path fault diagnosis. The diagnostic results show that the the simple probabilistic neuralnetwork training, rapid diagnosis and has good diagnostic accuracy.(4) Research on the gas path fault diagnosis technology based on clustering. Bycalculating the similarity, the clustering algorithm can distinguish between different failuremodes. The diagnostic results show that the clustering algorithm has a good diagnosticaccuracy and easy to understand, the algorithm can given the occurrence of certain types offailure probability.